The effect of the amount of prior knowledge on the long-term retrieval of information is examined. In two studies, seventh and eighth grade students of average or above average reading ability learned passages to simulate materials from history textbooks (factual), from reading textbooks (narrative), and from science textbooks (conceptual). Students studied passages for 45 seconds, 90 seconds, or 3 minutes depending upon whether the passage was at a high, medium, or low linkage level (i.e., prior knowledge level), respectively. During each test period, students said back what they could remember from the passage and the experimenter gave feedback. The criterion for learning was all but one proposition correct. Students free recalled the passages immediately after learning, and after four weeks. Students' retrospective reports were filled with links to prior knowledge, many of their errors appeared to include partial retrieval of propositions, and the prior knowledge structures surrounding higher prior knowledge passages had more alternate retrieval pathways. In a third study, the conditions of the recall test were identical to those already described except for changes to group testing and to a one-week interval. Although images did not affect recall, prior related knowledge did. (RL)